Osteoarthritis and Cartilage Vol. 15, Supplement C C147

values were greater than the respective median. The differencein treatment effect on baseline disease severity sub-group wasanalyzed using analysis of covariance (ANCOVA).Results: In total, 3235 patients comprised the ITT population ofthe pooled efficacy analysis. Out of these, 3214 patients wereincluded in the disease severity analysis [low (n = 923), medium(n =1456) and high (n=835)]. Mean OA pain at baseline was 79mm in the high disease severity, 65 mm in the medium diseaseseverity and 53 mm in the low disease severity subgroup. Thegreatest reduction in OA pain intensity with lumiracoxib andcelecoxib compared to placebo, was seen in the highest diseaseseverity group (lumiracoxib [-9.30mm; p≤0.001] celecoxib [-6.70mm; p=0.004]) as compared to the medium disease severitygroup (lumiracoxib [-5.48; p≤0.001], celecoxib [-4.76; p=0.006])and low disease severity group (lumiracoxib [-4.74 mm; p=0.012]celecoxib [-4.74 mm; p=0.030]). Similar results were seen forpatient’s global assessment of disease activity and the functionalstatus, assessed by WOMAC™ LK3.1 total score (Table 1).Conclusions: Lumiracoxib provides effective pain relief and im-provement of symptoms in all subgroups of OA patients definedby disease severity at baseline. However, there is a trend forgreatest efficacy in the subgroups of patients with highest dis-ease severity.

261

IS SELF-REPORTED IMPROVEMENT INOSTEOARTHRITIS PAIN AND DISABILITY REFLECTED INOBJECTIVE MEASURES?

G. Baltaci, V.B. TunayHacettepe University, Ankara, Turkey

Purpose: Osteoarthritis (OA) is perceived as a major publichealth problem, and today, various clinical outcome methods areused to see the improvement of the management for this condi-tion. The purpose of this study was to determine if self-reportedimprovements in pain and function correlate with improvement inobjective measures of disease in OA.Methods: Eighty female patients (mean age: 52.3 ± 8.7 years)with bilateral OA of the knee were assessed by some tests includ-ing [sociodemographics, body mass index (28.34± 4.88 kg/m2,and a self-reported questionnaire-Western Ontario and McMas-ter University Osteoarthritis Index (WOMAC) scores] and physi-cal [muscular strength by ISOMED 2000 isokinetic dynamometry,proprioception by determining deficit at the knee joint using byFunctional Squat System, Timed Get up and Go (TGG) for walk-ing duration, visual analog scale for measuring the intensity ofthe present pain, warmth, effusion] and clinical examination ofthe knees (Altman Grade II). Subjects received closed kineticchain in clinic and home-based structured exercise program 5days a week for 4 weeks, totally 20 sessions. All tests were donebefore and after the 4-week treatment period.Results: There were no significant correlation between pain atrest and night, during activity and proprioception deficit in left (r=0.05, r= -0.16, and r= -0.14; p>0.05) and right (r=0.22, r=0.07,and r=0.19; p>0.05) knee joint pre-treatment, respectively. Be-sides; there were found a correlation after treatment betweenpain at rest and night, with activity and proprioception deficit inleft (r= 0.32, r= -0.37, and r= -0.29; p<0.05) and right (r=0.53,r=0.39, and r=0.2; p<0.05) knee. The peak torque of isokineticmuscular strength and proprioception of participants were im-proved after rehabilitation. Of eighty eligible patients, WOMACscores were improved after immediate treatment in clinical-basedexercise group, 46.4% and 56.4% reported the improvementin pain and function, respectively. However, improvements inWOMAC scores were not associated with improvements in anyof the other measures (Left knee peak torque (PT), r= -0.28 and

right knee PT, r= -0.29; p>0.05 for all). There was a significantcorrelation between proprioception scores and TGG test eitherbefore (r= -0.31; p=0.016) (p<0.05) or after (r= -0.4141; p=0.001)(p<0.01) rehabilitation. The findings at 4th weeks after enroll-ment were that the quadriceps strength was slightly increasedbut the walking duration was decreased when compared withthe initial assessment. An exercise frequency of 20 sessions ina month was sufficient to improve muscle strength and walkingdistance.Conclusions: Seventy-three of participants reported significantimprovements in WOMAC pain and disability after treatment.Both closed kinetic chain and home-based exercise programproduce a significant improvement in strength and walking ability,especially in the first months. In the assessment of knee OA,muscle strength, proprioception, and walking ability can be usedto measure the effect of improvement of pain and disability aftertreatment programs.

262

DOES INCREASING OBESITY INCREASE THEPROGRESSION OF KNEE OSTEOARTHRITIS?

E. Vignon1, K. Brandt2, S.A. Mazzuca2, R. Buck3, B. Wyman4,M. Tengowski3, M-P. Hellio Le Graverand3

1Universite Claude Bernard, Lyon, France; 2Indiana UniversitySchool of Medicine, Indianapolis, IN; 3Pfizer Global Researchand Development, Ann Arbor, MI; 4Pfizer Global Research andDevelopment, Groton, CT

Purpose: Results of studies that have evaluated obesity as arisk factor for the progression of structural damage in knees withosteoarthritis (OA) are conflicting. However, researchers who de-sign clinical trials for structure-modifying drugs for OA (SMOADs)often consider “weighting” the study cohort with patients with ahigher body mass index (BMI), in the hope of enriching for pa-tients who will have more rapidly progressive OA. We askedwhether, among obese patients with knee OA, a higher BMI wasassociated with an increased rate of joint space narrowing (JSN).Methods: In 60 women with symptomatic OA (KLG2 or 3 in astanding AP radiograph) and a mean BMI (± SD) of 37.2 ±5.1, medial tibiofemoral JSN over 12 months were measured inmodified Lyon schuss (mLS) radiographs. We compared thesefindings to those in 81 age-matched, non-obese, asymptomatic,KLG0 women (BMI = 23.5 ± 2.3).Results: Among the OA patients, a higher BMI tended to beassociated with a higher KLG. Mean BMI was 36.0 ± 4.5 inKLG2 patients and 38.4 ± 5.4 in KLG3 patients (p=0.06). MeanJSN in the 81 controls was 0.02 ± 0.25 mm (p=0.51), whereasin the 30 KLG2 knees and 30 KLG3 knees, mean JSN was 0.12± 0.31 mm (p=0.08 vs. KLG0) and 0.32 ± 0.50 mm (p<0.0001vs. KLG0), respectively. KLG3 knees were more homogeneouswith respect to JSN than KLG2 knees (SRM = 0.64 vs. 0.39).BMI was not related to JSN in OA knees.Conclusions: Among these OA patients, with a BMI ≥ 30,there was no evidence that progressively higher BMIs wereaccompanied by a progressively increasing rate of JSN. Wewill determine whether a longer period of observation will yielddifferent results. In the mLS views, 12 months was sufficient todetect JSN in KLG2/3 knees, relative to knees of nonarthriticcontrols (p=0.0001). Because none of our OA patients had aBMI <30, this study cannot not address the question whetherJSN is more rapid in overweight or obese subjects than in kneeOA patients of normal weight. For an SMOAD trial, these datasuggest that recruitment of patients with a BMI much beyond 30will not enrich the sample in subjects who will exhibit more rapidJSN than would patients with a BMI of 30.